Skip to content

bhanu-kanamarlapudi/Covid_19-Michigan

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Covid-19 Michigan

Objective

The goal of this project was to accurately predict the number of daily COVID-19 cases across different counties in Michigan using machine learning models. Accurate predictions would allow public health officials to better allocate resources and plan pandemic response measures.

Data

For this project, I've used a variety of datasets to predict the number of COVID-19 cases for various counties in the state of Michigan. The datasets include

Preprocessing

The data was preprocessed to handle missing values, smooth noise in case count time series data, align different data sources by date, normalize features, and split into train/validation/test sets.

Models & Evaluation

Models

Several models were developed and evaluated:

  1. Multi-Linear Regression: to model linear relationships between weather, mobility, and cases
  2. MLP Regressor: a neural network to uncover complex non-linear relationships
  3. Gaussian Mixture Model Clustering: to cluster counties by case patterns

The models were trained on historical data and tuned using cross-validation.

Evaluation

The models were evaluated on held-out test data using root mean squared error (RMSE). The best model (MLP Regressor) obtained an average RMSE of 4 cases across all counties.

Summary

In summary, the machine learning models were able to leverage detailed data to make accurate county-level case predictions to assist public health decision-making for pandemic planning and response.